This invention relates to short pulse simultaneously Q-switched/cavity dumped and simultaneously super pulsed, Q-switched and cavity dumped CO2 lasers and more particularly to such lasers in material processing
It has become well appreciated in the laser machining industry that machined feature quality is improved as one utilizes shorter laser pulse widths and higher laser peak intensity in drilling holes. More specifically, the geometry of holes drilled with lasers become more consistent, and exhibits minimal recast layers and heat-affected zone around the holes as the laser pulses become shorter and their peak intensity becomes higher (XiangLi Chen and Xinbing Liu; Short Pulsed Laser Machining. How Short is Short Enough, J. Laser Applications, Vol. 11, No. 6, December 1999, which is incorporated herein by reference).
It is desirable to have the highest quality at the lowest cost but often one must choose a compromise. High-machined feature quality means low recast layer and heat-affected zone thickness, small surface roughness, accurate and stable machined dimensions. Low cost of ownership means a quick return on the investment made in the purchase of the laser machining equipment. Low cost of ownership also involves low maintenance, low operational costs, and high process speeds and yields in addition to low equipment cost. The choice of the laser parameters such as wavelength (IR, near IR, visible or UV lasers) and operational pulse format (milliseconds, microseconds, tenths of microseconds, nanoseconds, picosecond or femtosecond duration pulses) depends on the particular process, material design tolerance, as well as cost of ownership of the laser system.
Moving from lasers that function in the IR region (i.e. CO2) to the near IR (i.e. YAG or YLF), to the visible (i.e. doubled YAG or YLF), to the near UV (i.e. tripled YAG, YLF or excimer lasers), the trend is toward higher equipment cost in terms of dollar per laser average output power and lower average power output (which are disadvantages) while also having a trend toward higher power density (w/cm2) because of the ability to focus shorter wavelengths to smaller spot sizes (which is an advantage).
Moving toward shorter pulsed widths, the laser costs and the peak power per pulse and therefore power density (w/cm2) both tend to increase, while the average power output tends to decrease which results in the cost in terms of dollars per laser output power to increase.
The recast layer and heat-affected zone thickness are greatly reduced when using nanosecond pulses over millisecond and microsecond wide laser pulses (XiangLi Chen and Xinbing Liu; Short Pulsed Laser Machining: How Short is Short Enough, J. Laser Applications, Vol. 11, No. 6, December 1999). These improvements result from the higher laser beam intensity associated with the higher peak powers that are obtained with shorter laser pulses that utilize Q-switching, mode locking and other associated techniques and the fact that the pulse duration is shorter than the thermal diffusion time. For example, the typical thermal diffusion time for a 250 micron diameter hole is approximately 0.1 millisecond. In spite of the lower energy per pulse, high drilling speeds can still be cost effectively obtained because of the high pulse repetition rate obtained with these technologies. The high laser beam intensity provided by short laser pulses technology results in vaporization-dominated material removal rather than the melt-expulsion-dominated mechanisms using millisecond wide laser pulses. It is also known that shorter pulse width yield more limited heat diffusion into the surrounding material during the laser pulse. Hole-to-hole dimensional stability is also improved because the hole is drilled by the material being nibbled away by tens to hundreds of laser pulses of smaller pulse energy but occurring at a high pulse repetition frequency rather than by a few high-energy pulses. For the same reason, thermal and mechanical shocks from nanosecond pulses are also reduced compared with millisecond pulses. These advantageous effects obtained with nanosecond laser pulses have been detected by observing fewer micocracks occurring when holes were drilled in brittle materials such as ceramic and glass when utilizing nanosecond laser pulses.
When the intensity is further increased through laser mode locking techniques to get down to the subnanosecond pulse width (i.e. picoseconds and femtosecond region), additional reductions in the recast and heat-affected zones are observed. Since a typical electron energy transfer time is in the order of several picoseconds, femtosecond laser pulse energy is deposited before any significant electron energy transfer occurs within the skin depth of the material. This forms a plasma that eventually explodes and evaporates the material leaving almost no melt or heat-affected zone. Due to the small energy per pulse (xcx9c1 mJ), any shock that is generated is weak resulting in no microcracks even in brittle ceramic alumna material. Femtosecond pulses are not presently obtainable with CO2 lasers due to the narrow gain of the laser line. Femtosecond pulses are presently obtainable with solid-state lasers.
For the same total irradiated laser energy, femtosecond pulses remove two to three times more material than the nanosecond pulses. However, even xe2x80x9cheroxe2x80x9d type, one of a kind experimental, state of the art laser research and development systems that operate in the femtosecond range deliver only several watts of average power, while nanosecond lasers yield one or two order of magnitude higher power output. Consequently, femtosecond lasers are still too low in average power to deliver the required processing speeds for most commercial applications. It has been reported (XiangLi Chen and Xinbing Liu; Short Pulsed Laser Machining: How Short is Short Enough, J. Laser Applications, Vol. 11, No. 6, December 1999) that a 1W femtosecond laser requires more than a minute to drill a 1.0 mm deep hole of 0.1 mm diameter. Present femtosecond lasers have such high cost that their use is cost effective for only special high value applications that unfortunately have relative low unit volume market potential. For example, Lawerance Livermore National Lab has made use of the fact that femtosecond laser pulse energy is deposited essentially with no thermal transfer to cut and shape highly sensitive explosive materials without denotation.
It is well known that the trend for optical absorption in metals as a function of wavelength is toward lower absorption with increasing wavelengths as shown in FIG. 1. Consequently, the near IR, visible and ultra violet wavelength regions are most effective in machining most metals. This advantage does not exist in plastic material. The data contained in FIG. 1 is not relevant once a plasma is initiated on the metal surface because all of the laser energy is absorbed in the plasma, which in turn imparts the energy to the material. Once the plasma is initiated, the absorption as a function of wavelength variation for metals becomes essentially flat. Consequently, one can paint the surface of the metal for greater absorption at longer wavelengths and the higher absorption advantage of shorter laser wavelengths is effectively eliminated.
The electronics industry has needs to shrink the size of semiconductor and hybrid packages, and greatly increase the density of printed circuit boards because of the market desire for smaller cellular phones, paging systems, digital cameras, lap top and hand held computers, etc. These needs have resulted in interest in the use of lasers to form small vertical layer-to-layer electrical paths (via) in printed circuit boards. The short pulse CO2 laser is particularly attractive for drilling via holes in printed circuit boards because of 1) the high absorption of the printed circuit board or hybrid circuits resin or ceramic material at the CO2 wavelength when compared to YAG or YLF lasers which operate in the near IR and in the visible and UV wavelength regions with harmonic generating technique; 2) the lower cost per watts associated with CO2 lasers when compared to YAG lasers, 3) and because of the high reflectivity of copper at CO2 wavelengths, which enables CO2 laser via hole drilling equipment to drill through the resin layer down to the copper layer where the drilling is stopped because of the high reflectivity of the copper interconnect material at the CO2 laser wavelengths. These are called xe2x80x9cblind via,xe2x80x9d which connect the outer layer of a circuit to the underlying inner layer within the multi layer board. The major disadvantages of CO2 lasers in via hole drilling is the larger spot size obtainable with its 10.6 micron wavelength when compared to shorter wavelength laser. Another disadvantage is that pulse widths below several nanosecond are difficult to obtain with CO2 lasers. The major advantages of CO2 Q-switched lasers are: they offer lower cost per watt of laser output when compared with solid state lasers, the higher absorption of their radiation by resin and ceramic board materials, their ability to operate at high PRF, their ability to generate substantial output power under Q-switched operation, and their ability to stop drilling when the radiation gets to the copper layer.
The advantages of drilling via holes in printed circuit boards with laser systems have enabled laser systems to capture 70% of the via hole machine drilling market in 1999 (David Moser; Laser Tools For Via Formation, Industrial Laser Solutions, p. 35, May, 2000, which is incorporated herein by reference), with the remaining 25% of the market held by photo-via and the remaining 5% by other techniques, such as mechanical drilling, punch and plasma etching.
The upper CO2 laser transition level has a relatively long decay rate for storing larger than normal population inversion (385 torrxe2x88x921 secxe2x88x921 at 300 K and approximately 1300 torrxe2x88x921 secxe2x88x921 at 500 K). The lower CO2 laser levels for both the 9.4 and 10.4 transitions are approximately an order of magnitude faster so a large population inversion between the lower levels and the upper level can be easily maintained. Laser mediums that have transitions with long lifetime upper energy levels are good candidates for application of Q-switched techniques (A. E. Siegrnan; Lasers, Chapt. 26, University Science Books, 1986, which is incorporated herein by reference). The long lifetime of the upper levels store energy by building up a higher than normal population with respect to the lower laser level. Consequently, CO2 lasers are good candidates for performing Q-switching (G. W. Flynn et al; Progress and Applications of Q-switching Techniques Using Molecular Gas Lasers, IEEE J. Quant. Electronics, Vol. QE-2, p. 378-381, Sept. 1966, which is incorporated herein by reference). Consequently, CO2 lasers are good candidates for simultaneously Q-switching and cavity dumping in order to obtain shorter pulse widths than obtainable from just Q-switching.
Q-switching is a widely used technique in which a larger than normal population inversion is created within a laser medium by initially providing for a large loss within the feedback cavity. After a large inversion is obtained, one quickly removes the large optical loss within the feedback cavity, thereby quickly switching the cavity Q back to its usual large value (i.e. low loss value). This results in a very short intense burst of laser output, which dumps all the excess population inversion into the short laser pulse (A. E. Siegman; Lasers, Chapt. 26, University Science Books, 1986).
FIG. 47A illustrates the time dependent variation of the losses within the feedback cavity that can be obtained with either a rotating feedback mirror, an electro-optics modulator (i.e. switch) or with an acousto-optics switch inserted in the lasers feedback cavity under continuous pumping condition shown in FIG. 47B. FIG. 47A also illustrates the time dependent gain variation experience by the continuously excited laser under the internal optical cavity loss variations illustrated. The gain is allowed to rise for an optimum time of about one or two population decay time of the upper CO2 laser level. At such an optimum time, the cavity loss is switched from a high loss to the normal loss condition (or the Q of the cavity is switched from a low to a high value condition) by applying a high voltage pulse to the electro-optic modulator (EOM) as shown in FIG. 47B. Since the gain greatly exceeds the losses at this point, laser oscillations by stimulated emission begins with the output building up exponentially, resulting in the emission of a giant laser output pulse whose peak power is hundreds of times larger than the continuous power of the laser. The pulse has a long tail, which will eventually decay down to the lasers"" CW power level as long as the gain exceeds the feedback cavity loss. This tail is detrimental to most hole drilling process. The use of simultaneously Q-switched and cavity dumped process as revealed in this invention will eliminate this long pulse tail problem associated with Q-switching. When the high loss cavity condition is again switched on, the laser action stops and the described dynamic process is repeated.
To the present time, Q-switched and cavity dumped CO2 lasers have not found extensive commercial application, as have Q-switched solid-state lasers (whose upper state life times are measured in seconds instead of tenths of seconds as for the CO2 laser). Nearly all of the Q-switched CO2 laser applications to date have addressed predominately military and scientific applications. Some of the reasons for the lack of interest in commercial CO2 Q-switched and cavity dumped lasers are high cost of the electro-optic crystal (namely CdTe), limited suppliers for the electro-optic (EO) crystals, large performance variation between different optical paths within an EO crystal and large performance variation between different crystals. There is also difficulty in obtaining good anti reflection thin-film coatings on CdTe crystals. In addition, acousto optic modulators in the IR have higher attenuation and poorer extinction performance than in the visible region, as well as larger thermal distortion and poorer reliability. Based upon these EO crystal limitations CO2 lasers were considered to have poorer reliability than the Q-switched solid state laser which was mostly attributed to the CdTe crystals. Consequently, superpulse operation, or mechanically Q-switched CO2 lasers and TEA CO2 laser techniques have been utilized to satisfy most short pulse CO2 laser industrial needs to date (A. J. DeMaria; Review of CW High Power CO2 Lasers, Proceedings of the IEEE, pg. 731-748, June 1973, which is incorporated herein by reference).
For these reasons, techniques such as gated CW and super pulse, along with external cavity acousto-optic deflection of either CW or super pulsed lasers into a aperture have been predominately utilized to date with CO2 lasers to obtain IR laser pulses for industrial applications. Each of these techniques are deficient when compared with CO2 Q-switching or simultaneous Q-switching and cavity dumped operation in one or more of the following: 1) longer pulse widths with slower rise time, 2) lower pulse repetition frequencies (PRF), 3) lower over all laser efficiencies, 4) long duration tails associated with the pulses and 5) lower peak powers. TEA lasers have also been used to date, but they suffer from higher time jitter from pulse to pulse, higher pulsed voltage requirements to energize the laser along with associate acoustic shock noise from the discharge, low pulse repetition rates and non-sealed off laser operation which requires gas flow.
Thus it is desirable to make both the Q-switched and Q-switched and cavity dumped CO2 laser lower in cost, more reliable, to enable the cost effective utilization of the present state of the art of CdTe electro-optics crystal technology without sacrificing Q-switching performance, and obtaining higher peak power and shorter pulses by simultaneously utilizing super pulse and Q-switching techniques as well as cavity dumping techniques. It is desirable to make simultaneously Q-switched and cavity dumped CO2 lasers commercially practical because pulse widths of 7-20 nsec are obtainable while Q-switched CO2 lasers provide pulse widths of 100-150 nsec. Shorter pulse have advantages for numerous hole drilling applications, especially for via hole drilling of printed circuit board and for laser marking of stressed glass containers holding a vacuum or partial vacuum or a pressure higher than ambient pressure such as automobile headlights, flat panel displays, cathode ray tubes for TVs and computers, street lights, light bulbs stressed plate glass in automobiles or pressured glass or plastic containers containing soft drinks, beer, etc.
FIG. 2 illustrates a block diagram of a laser material processing system. The system includes the laser head and its power unit, which may or may not have an internal controller. An RF power unit is preferred. The RF power unit can be operated CW or in the super pulsed mode. The super pulsed mode is used to obtain increased peak power laser pulses. The laser head and its power unit and controller are usually provided by a laser supplier, while the controller for the XY scanning system, the scanners, the keyboard, the optical shutter and a display unit are usually the responsibility of the original equipment manufacturer. The original equipment manufacturer (OEM) controller commands the scanning system and the display unit and sends signals to the laser controller, which in turn commands the laser head. If the laser is liquid cooled, a chiller is required which either the laser manufacturer or the systems OEM can supply. Usually, the OEM chooses to supply the chiller. Laser beam shaping optics are usually required between the laser head and the scanners. Either the laser manufacturer or the OEM system manufacturer can supply the laser beam shaping optics. This overview block diagram is essentially identical to a block diagram used to describe laser engraving, marking, cutting and drilling systems for desk top manufacturing type applications with the software being basically the differentiating portion of the system. The system OEM normally is responsible for the optical scanner, the system controller and its software and the displays.
The OEM system controller tells the XY optical scanning system where to point and informs the laser head through the controller within the laser""s head power unit when to turn on or off and how much power is to be emitted. The OEM system controller also monitors and supervisors the chiller, and displays the desired information on the display unit to the system operator who usually enters commands through the keyboard that address the system controller. The system controllers and the laser power unit controller also perform appropriate diagnostics to protect the system from inadequate cooling, RF impedance mismatch between the laser discharge and the RF power supply, and safety features such opening and closing the systems optical shutter, etc.
FIG. 3 illustrates the modifications to FIG. 2 for the case when a simultaneously Q-switched, cavity dumped laser is utilized in the material processing system. In addition to commanding the laser power supply, the system controller performs calculations utilizing the input from the operator provided through the keyboard and issues commands regarding the laser modulation format (i.e. gated output or super pulse output for example, the timing of the Q-switched or the simultaneously Q-switched, cavity dumped laser laser pulse along with pulse duration and repetition frequency, etc.) and monitoring the status of the laser head and its power supply as well as the chiller. The system controller also issues commands (and may receive signals) from the Q-switched power module. The system controller receives signal from an operator through a keyboard and commands as well as monitors the status of the optical shutter, which can be inserted either before or after the optical scanners. In some cases, the optical shutter is specified for inclusion at the direct exit of the laser beam out of the laser housing. If the shutter is included as part of the laser housing, the laser manufacturer supplies the optical shutter and its circuitry. The status of the system is displayed to the operator of the keyboard by an appropriate display unit. The Q-switching module of FIG. 3 is in principle the same for either a solid state or gas laser system with the major difference being the use of a different electro-optical crystal.
In addition to utilizing Q-switched or a simultaneously Q-switched, cavity dumped lasers and even shorter pulsed laser systems, such as mode locked laser systems for hole drilling applications, the laser system of FIG. 3 can also be utilized to mark, encode or drill stressed glass vessels or structures as well as to perforate or punch holes in paper without charring. The advantage of utilizing simultaneously Q-switched cavity dumped lasers over Q-switched lasers to mark or encode stressed glass containers, which have a pressure difference between the inside and outside surfaces of the containers, arises because they can provide an order of magnitude shorter laser pulse widths. The use of simultaneously super pulsed Q-switched and cavity dumped lasers for marking and machining stressed glass containers has not been appreciated nor recognized because such short pulse CO2 laser systems for such applications have not been presently commercially available. Such containers include, for example, sealed glass automotive headlights, streetlights, cathode ray tubes, flat panel displays and beer, soda, and champagne bottles. Tempered glass surfaces of safety glass doors, windows, and automotive side windows are also good candidates for laser marking or encoding with short laser pulses because the generation of microcracks in brittle materials such as glass and ceramic materials are greatly reduced by Q-switched or cavity dumped laser pulses. If longer pulsed laser radiation is used to mark such stress containers and glass surfaces, micro cracks are created at the location where the laser marks or encodes the glass. These microcracks become enlarged and propagate with time under the stress load that the brittle material is subjected. CO2 laser radiation is strongly absorbed by glass and ceramics so they are the laser of choice for such applications. Because of their size, power, cost and processing speed, CO2 lasers are preferred for non-metal processing of materials. UV radiations are also absorbed by glass material and are considered alternate lasers for such applications, but at higher cost and slower processing speeds.
The high laser beam intensity provided by short pulse laser technology results in the vaporization-dominated material removal rather than the melt-expulsion-dominated mechanisms using longer duration pulses. Thermal and mechanical shocks are reduced with the short laser pulse system of FIG. 3 when compared with longer pulse systems of FIG. 2. Consequently, micro cracks are greatly reduced under laser marking or encoding with short pulse lasers. Simultaneously Q-switched cavity dumped laser output pulses do not have the long tails associated with Q-switched laser pulses. Consequently, the formation of micro cracks at the glass location, which is marked or encoded is reduced to a larger extent then with the Q-switched CO2 laser. The application of the laser system of FIG. 3 thereby opens up the market of direct marking or encoding on stressed glass containers and structures. Currently ink jets or other similar devices are used to mark or encode such glass containers and structures. Inkjets have well known disadvantages over laser marking/encoding system. Some of these disadvantages are their mark is not permanent and can rub off through handling, and exposure to the environment, the inks and solvents are consumables and recurring costs can be high, the inks and solvents are toxics and dirty up the factory environment and the down time of inkjet marking systems is high which adds to their operating costs. The major advantage of inkjet marking systems for this application is low initial capital cost.
The drilling of numerous small holes in paper or plastic parts without charring the edges of the paper or plastic material is desired in many industries. Some examples are in the tobacco filtration, and in the banking and billing industries for perforating checks and other financial documents. In the past TEA lasers, which suffer from low pulse repetition frequencies, were used for these applications. It has not been appreciated that Q-switched and simultaneously Q-switched and cavity dumped lasers lasers can be utilized to perforate such materials. If higher energies are required than those available with a single sealed-off Q-switched cavity dumped laser, then a laser amplifier can be used to increase the pulse energy of the Q-switched cavity dumped laser. Q-switched lasers and simultaneously Q-switched cavity dumped lasers have pulse repetition frequencies of greater than 100 kHz while TEA lasers have an upper practical limit of PRF of 500 Hz.
Background on Cavity Dumping
Simultaneous Q-switched cavity dumped techniques can be used in CO2 lasers to obtain pulse widths of approximately 7 to 20 nanoseconds depending upon the length of the laser feedback cavity. Cavity dumping is a technique in which the output coupling out of the laser cavity is suddenly increased to a very large value. This is as if one of the end laser mirrors had been removed so that all of the circulating energy within the cavity is xe2x80x9cdumpedxe2x80x9d into an output pulse. For perfect dumping, the output pulse will be exactly one cavity round-trip time T in duration, and contains all of the optical energy contained within the cavity. FIG. 4 shows an apparatus by which this can be accomplished.
The laser cavity comprising mirrors 406, 408 of FIG. 4 contains a polarizing beam splitter 405, which makes the cavity 406, 408 oscillate normally with a given polarization, for example, with a vertically oriented linear polarization as illustrated in FIG. 4. In order to dump the cavity energy, the voltage across the electro-optic modulator 140 (e.g. a Pockels cell) is suddenly switched (in a time short compared to T, or typically a few nanoseconds) to a value which makes the transparent electro-optic crystal 142 become birefringent, with a value corresponding to a quarter-wave plate for single pass, or a half-wave plate for double pass. The laser radiation that was linearly polarized in the plane of the page, becomes circularly polarized as it passes through the crystal 142 going to the right and then coming back through the crystal 142 to the left, after being reflected off mirror 408 has its polarization converted into a polarization perpendicular to the plane of the page after coming back out of the crystal 142. All the energy in this polarization coming back to the polarizing beam-splitter 405 (which can either be an specially coated dielectric plate, as shown, or a polarizing prism, such as a Glan-Thompson prism) is then reflected (e.g. dumped) out of the cavity 406, 408 as shown at 407. The pulsed radiation 407 has a pulse width essentially equal to the time required for light to make one round trip between the mirrors 406 and 408. As a practical matter, a fixed quarter-wave plate (not shown) is often added to the EOM 140, and the EOM 140 is then initially biased to a fixed voltage, typically several thousand volts, required to cancel the fixed quarter-wave plate. Since it is easier to short out or xe2x80x9ccrowbarxe2x80x9d the EOM voltage from a high initial value down to zero volts in a few nanoseconds (nsec) than it is to switch the same voltage from zero volts up to the necessary high voltage value in the same length of time, cavity dumping may be accomplished by suddenly switching off the voltage across the EOM, leaving only the fixed quarter-wave plate. This approach is not well suited with CdTe EOM crystals due to traces of impurities which adversely affect the performance of these crystals in electro-optical modulator applications under prolonged exposure to high voltage. In addition, the temperature changes in the quarter wave plate causes variations of the polarization of the radiation propagating through the 1/4 xcex plate.
Consider as an example, a CW laser oscillator running in steady-state with an approximate 5% output coupling mirror (R=95%), the circulating intensity inside the laser cavity is then 20 times as large as the CW output intensity from the laser. If this circulating intensity is suddenly cavity dumped, the peak power output during the dumped pulse can be up to 20 times as large as the average or CW power output from the laser. By reducing the output coupling essentially to zero, one can make both the circulating radiation intensity and thus the xe2x80x9cdumpablexe2x80x9d peak power still larger. It has been found that coated metal mirrors are well suited for highly reflecting feedback mirrors without suffering mirror damage under the high intensity radiation within the laser cavity.
If one further allows the intensity inside the cavity to build back up again, and one then again dumps the laser cavity, using repetitive cavity dumping, one can obtain most of the available power output from the laser medium as output in the form of repeated pulses, which have substantially higher peak power than the average power from the laser. With proper choice of repetition frequency, the average power in the dumped output can approach the full average power available with optimum coupling in CW operation; but the higher peak powers can make this energy much more effective in drilling, cutting, welding, and other nonlinear laser processes.